A centrifugal pump performance evaluation method and system based on cloud computing

Abstract

The application provides a centrifugal pump performance evaluation method and system based on cloud computing, relates to the technical field of centrifugal pump performance evaluation, can dynamically reflect the influence of slurry characteristic change on pump performance by introducing and analyzing fluid resistance influence quantity, thereby avoiding misjudgment of temporary performance fluctuation caused by slurry viscosity change as pump body wear. Meanwhile, when the pump body actually wears slowly, even if the slurry characteristics temporarily cover up the performance decline, the application can also identify the real physical damage in time through the low fluctuation state of the fluid resistance influence quantity and the continuous exceeding of the performance deviation from the preset range, thereby avoiding potential failure risks. Therefore, the application significantly improves the accuracy and reliability of the centrifugal pump performance evaluation, provides a more scientific and refined basis for centrifugal pump maintenance decision in chemical production, effectively reduces the operation cost, and guarantees the continuity and safety of production.

Description

Technical Field

[0001] This application relates to the field of centrifugal pump performance evaluation technology, and more specifically, to a cloud computing-based centrifugal pump performance evaluation method and system. Background Technology

[0002] In large-scale chemical production facilities, centrifugal pumps are key equipment for maintaining the continuous and stable operation of the process. To ensure production continuity and reduce energy consumption, factories typically deploy cloud-based centrifugal pump performance evaluation systems. This system collects real-time operating data from the centrifugal pumps using sensors and uploads it to a cloud computing platform. The cloud platform has pre-established an ideal performance model based on clean water as the medium. By comparing the real-time operating data with the model, the system determines that the pump performance has deteriorated and issues a warning if a deviation is found between the actual and ideal performance.

[0003] However, in actual chemical production, centrifugal pumps typically transport chemical slurries containing solid particles, whose physical properties (such as viscosity and solid content) fluctuate with each production batch. This variation in slurry characteristics has two effects on pump performance: first, the physical wear of pump components (such as impellers and pump casings) by hard particles in the slurry leads to an irreversible and permanent decline in the pump's hydraulic performance; second, temporary performance fluctuations caused by changes in slurry viscosity. For example, when the slurry is viscous, fluid resistance increases, and the pump outlet pressure and flow rate decrease in the short term, while they recover when the slurry is thin.

[0004] Existing cloud-based assessment systems cannot effectively distinguish between these two fundamentally different performance changes. Because their models are static and lack the ability to understand the dynamic changes in the fluid medium, the system attributes all observed performance degradation to physical damage to the pump. This confusion leads to incorrect maintenance decisions. For example, when viscous slurry causes performance parameter declines, the system may falsely report pump wear, resulting in unnecessary downtime and wasted resources. Conversely, when slow wear actually occurs in the pump, if a thin slurry is being transported, the pump's output performance may temporarily remain within acceptable limits, masking the true physical damage and causing the system to fail to issue timely warnings, increasing the potential risk of failure. Summary of the Invention

[0005] This application provides a cloud computing-based centrifugal pump performance evaluation method and system, which aims to solve the problem that existing centrifugal pump performance evaluation systems cannot effectively distinguish between performance fluctuations caused by changes in slurry characteristics and performance degradation caused by physical wear of the pump body.

[0006] On the one hand, this application provides a cloud computing-based centrifugal pump performance evaluation method, including:

[0007] Based on the centrifugal pump's operating data, calculate the actual motor input power, actual head, and actual efficiency. The operating data includes the centrifugal pump's inlet pressure, outlet pressure, liquid flow rate, motor input current, and motor input voltage.

[0008] Based on the preset clean water benchmark performance data, obtain the theoretical clean water head and theoretical clean water efficiency corresponding to the liquid flow rate, and calculate the theoretical clean water motor input power based on the theoretical clean water head and the theoretical clean water efficiency.

[0009] Calculate the influence of fluid resistance based on the actual motor input power and the theoretical clean water motor input power;

[0010] Based on the fluid resistance influence, the theoretical clean water head and theoretical clean water efficiency are corrected to generate the target theoretical clean water head and target theoretical clean water efficiency.

[0011] The actual head and actual efficiency are compared with the target theoretical head and target theoretical water efficiency to obtain the comparison results. Based on the comparison results and the changing trend of the fluid resistance influence, the performance fluctuations caused by changes in slurry characteristics and the performance degradation caused by physical wear of the pump body are distinguished.

[0012] On the other hand, this application also provides a cloud-based centrifugal pump performance evaluation system, which includes:

[0013] The power calculation module is used to calculate the actual motor input power, actual head, and actual efficiency based on the centrifugal pump's operating data. The operating data includes the centrifugal pump's inlet pressure, outlet pressure, liquid flow rate, motor input current, and motor input voltage.

[0014] The benchmark establishment module is used to obtain the theoretical clean water head and theoretical clean water efficiency corresponding to the liquid flow rate based on the preset clean water benchmark performance data, and to calculate the theoretical clean water motor input power based on the theoretical clean water head and the theoretical clean water efficiency.

[0015] The resistance influence calculation module is used to calculate the fluid resistance influence based on the actual motor input power and the theoretical clean water motor input power.

[0016] The standard correction module is used to correct the theoretical clean water head and theoretical clean water efficiency based on the fluid resistance influence amount, and generate the target theoretical clean water head and target theoretical clean water efficiency.

[0017] The performance diagnostic module is used to compare the actual head and actual efficiency with the target theoretical clean water head and target theoretical clean water efficiency to obtain the comparison results. Based on the comparison results and the changing trend of the fluid resistance influence, it distinguishes between performance fluctuations caused by changes in slurry characteristics and performance degradation caused by physical wear of the pump body.

[0018] This application relates to a cloud-based centrifugal pump performance evaluation method and system. By acquiring the centrifugal pump's operating data, it calculates the actual motor input power, actual head, and actual efficiency. Based on preset clean water benchmark performance data, it obtains the theoretical clean water head, theoretical clean water efficiency, and theoretical clean water motor input power. In addition, this application innovatively introduces the calculation of the fluid resistance influence, which is obtained by the difference between the actual motor input power and the theoretical clean water motor input power, effectively quantifying the impact of slurry characteristic changes on pump performance. Subsequently, the fluid resistance influence is used to correct the theoretical clean water head and theoretical clean water efficiency, generating target theoretical clean water head and target theoretical clean water efficiency that better reflect actual operating conditions. Finally, by comparing the actual performance parameters with the corrected target theoretical performance parameters and combining the changing trend of the fluid resistance influence, this application can accurately distinguish between performance fluctuations caused by changes in slurry characteristics and performance degradation caused by physical wear of the pump body.

[0019] Compared to existing technologies, this application overcomes the technical limitations of traditional cloud-based assessment systems, which cannot effectively distinguish between changes in slurry characteristics and physical wear of the pump body. Existing systems, due to their static models' inability to understand dynamic changes in the fluid medium, often attribute all performance degradation to physical damage to the pump, leading to false alarms and unnecessary downtime. This application, by introducing and analyzing the influence of fluid resistance, can dynamically reflect the impact of changes in slurry characteristics on pump performance, thereby avoiding misjudging temporary performance fluctuations caused by changes in slurry viscosity as pump wear. Furthermore, when slow wear actually occurs in the pump body, even if slurry characteristics temporarily mask the performance degradation, this application can promptly identify actual physical damage through the low fluctuation of the fluid resistance influence and the continuous deviation of performance from the preset range, avoiding potential failure risks. Therefore, this application significantly improves the accuracy and reliability of centrifugal pump performance assessment, providing a more scientific and refined basis for centrifugal pump maintenance decisions in chemical production, effectively reducing operating costs, and ensuring the continuity and safety of production. Attached Figure Description

[0020] To illustrate this application more clearly, the accompanying drawings used in the embodiments will be briefly described below. Obviously, those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0021] Figure 1 An exemplary flowchart of a cloud-based centrifugal pump performance evaluation method is shown.

[0022] Figure 2 An exemplary schematic diagram of a cloud-based centrifugal pump performance evaluation system is shown.

[0023] Figure reference numerals: 100, Cloud-based centrifugal pump performance evaluation system; 10, Power calculation module; 20, Benchmark establishment module; 30, Resistance influence calculation module; 40, Standard correction module; 50, Performance diagnosis module. Detailed Implementation

[0024] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0025] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0026] Traditional centrifugal pump performance evaluation systems often fail to effectively distinguish between performance fluctuations caused by changes in slurry characteristics and performance degradation due to physical wear and tear on the pump body when handling chemical slurries containing solid particles. This confusion can lead to incorrect maintenance decisions, such as unnecessary downtime or failure to provide timely warnings of actual physical damage.

[0027] like Figure 1 The diagram illustrates an exemplary flowchart of a cloud-based centrifugal pump performance evaluation method. This application proposes a cloud-based centrifugal pump performance evaluation method, comprising:

[0028] S10, Calculate the actual motor input power, actual head, and actual efficiency based on the centrifugal pump's operating data, wherein the operating data includes the centrifugal pump's inlet pressure, outlet pressure, liquid flow rate, motor input current, and motor input voltage.

[0029] Centrifugal pump operating data refers to various physical parameters collected in real time by sensors and other equipment during the actual operation of the centrifugal pump, such as inlet pressure, outlet pressure, liquid flow rate, motor input current, and motor input voltage. These data form the basis for evaluating centrifugal pump performance. Actual motor input power refers to the electrical power input by the motor during actual operation, reflecting the energy consumed by the pump in overcoming various resistances (including fluid resistance and mechanical resistance). Actual head refers to the energy increment gained per unit weight of liquid after passing through the pump during actual operation, usually expressed as the height of the liquid column, and is an important indicator of the pump's ability to lift liquid. Actual efficiency is the ratio of the effective power of the centrifugal pump to the actual motor input power, reflecting the efficiency with which the pump converts electrical energy into liquid energy.

[0030] S20: Based on the preset clean water reference performance data, obtain the theoretical clean water head and theoretical clean water efficiency corresponding to the liquid flow rate, and calculate the theoretical clean water motor input power based on the theoretical clean water head and the theoretical clean water efficiency.

[0031] S30, calculate the influence of fluid resistance based on the actual motor input power and the theoretical clean water motor input power;

[0032] S40, based on the fluid resistance influence, correct the theoretical clean water head and theoretical clean water efficiency to generate the target theoretical clean water head and target theoretical clean water efficiency;

[0033] S50, compare the actual head and actual efficiency with the target theoretical head and target theoretical efficiency of clean water to obtain the comparison results, and based on the comparison results and the changing trend of the fluid resistance influence, distinguish between performance fluctuations caused by changes in slurry characteristics and performance degradation caused by physical wear of the pump body.

[0034] This application, by introducing and analyzing the influence of fluid resistance, can effectively distinguish between performance fluctuations caused by changes in slurry characteristics and performance degradation caused by physical wear of the pump body. This avoids misjudgments by traditional evaluation systems under complex operating conditions and improves the accuracy of evaluation and the reliability of maintenance decisions.

[0035] To better understand the method proposed in this application, some key terms are explained first. The theoretical clean water head and theoretical clean water efficiency are the ideal head and efficiency that a centrifugal pump should achieve at a specific liquid flow rate, based on clean water benchmark performance data. Fluid resistance influence is a key concept introduced in this application; it quantifies the impact of the difference in fluid resistance between the actual operating medium (such as slurry) and the clean water medium on pump performance. The target theoretical clean water head and target theoretical clean water efficiency are obtained by correcting for the theoretical clean water head and theoretical clean water efficiency using the fluid resistance influence, thus more closely approximating the ideal performance parameters of the pump under actual operating conditions.

[0036] The core of the cloud computing-based centrifugal pump performance evaluation method proposed in this application lies in distinguishing performance changes caused by different reasons through in-depth analysis of operating data.

[0037] First, it is necessary to acquire the centrifugal pump's operating data. This data forms the basis of the evaluation and is typically collected in real time by installing appropriate sensors at the pump's inlet, outlet, and motor. For example, pressure sensors are used to measure inlet and outlet pressures, flow meters measure liquid flow rate, and current and voltage transformers measure motor input current and voltage. These sensors convert the acquired analog signals into digital signals, which are then uploaded to a cloud server for storage and processing via a data acquisition unit.

[0038] Secondly, based on the acquired operating data, the actual motor input power, actual head, and actual efficiency of the centrifugal pump are calculated. The actual motor input power can be calculated from the motor input current and motor input voltage. For example, for a three-phase AC motor, the actual motor input power can be calculated using the formula... Calculation, where Line voltage, For line current, This refers to the power factor. The actual head can be calculated using parameters such as inlet pressure, outlet pressure, liquid density, and pump installation height. ,in Due to export pressure, For inlet pressure, For the density of the liquid, It is the acceleration due to gravity. This represents the pump's installation height difference. The actual efficiency can be calculated as the ratio of actual hydraulic power to actual motor input power, where actual hydraulic power can be calculated from the liquid flow rate, actual head, and liquid density.

[0039] Next, based on preset clean water reference performance data, the theoretical clean water head and theoretical clean water efficiency corresponding to the liquid flow rate are obtained, and the theoretical clean water motor input power is calculated. Clean water reference performance data is typically stored in a cloud database in the form of tables or curves, recording the theoretical head and theoretical efficiency of the centrifugal pump at different liquid flow rates when conveying clean water. In practical applications, the theoretical clean water head and theoretical clean water efficiency corresponding to the current liquid flow rate can be obtained from this data using table lookup or interpolation methods (such as linear interpolation). For example, if the clean water reference performance data is stored in discrete point form, when the actual liquid flow rate is between two discrete points, the corresponding theoretical clean water head and theoretical clean water efficiency can be calculated using linear interpolation. After obtaining the theoretical clean water head and theoretical clean water efficiency, the theoretical clean water hydraulic power can be calculated based on the liquid flow rate and liquid density, and then the theoretical clean water motor input power can be calculated based on the preset clean water efficiency.

[0040] Then, based on the actual motor input power and the theoretical clean water motor input power, the influence of fluid resistance is calculated. The influence of fluid resistance is a key indicator measuring the difference in fluid resistance between the actual conveyed medium (such as slurry) and the clean water medium. One calculation method is to directly use the difference between the actual motor input power and the theoretical clean water motor input power as the influence of fluid resistance. For example, when a centrifugal pump conveys slurry, the viscosity or solid content of the slurry is higher than that of clean water, leading to increased fluid resistance. This results in the actual motor input power being higher than the theoretical clean water motor input power; the difference reflects the impact of fluid resistance on power.

[0041] Subsequently, based on the influence of fluid resistance, the theoretical clean water head and theoretical clean water efficiency are corrected to generate the target theoretical clean water head and target theoretical clean water efficiency. The influence of fluid resistance reflects the difference in fluid characteristics between the actual medium and the clean water medium, which directly affects the pump's head and efficiency. Therefore, it is necessary to use the influence of fluid resistance to correct the theoretical clean water head and theoretical clean water efficiency to make them closer to the ideal performance of the pump under actual operating conditions. For example, when the influence of fluid resistance is positive, it indicates that the fluid resistance of the actual medium is relatively large. In this case, the theoretical clean water head can be appropriately reduced and the theoretical clean water efficiency can be appropriately adjusted according to the preset correction coefficient or model to obtain the target theoretical clean water head and target theoretical clean water efficiency.

[0042] Finally, the actual head and efficiency are compared with the target theoretical head and efficiency for clean water, yielding comparison results. Based on these results and the changing trends of the fluid resistance influence, the performance fluctuations caused by changes in slurry characteristics are distinguished from performance degradation caused by physical wear of the pump body. The comparison results can be expressed as performance deviations, i.e., the differences between actual performance parameters and target theoretical performance parameters. By analyzing the magnitude and trend of performance deviations, combined with the changing trends of the fluid resistance influence, the causes of pump performance degradation can be diagnosed. For example, if the performance deviation consistently exceeds the preset range, and the fluid resistance influence remains low, it may indicate physical wear of the pump body; if the performance deviation and the fluid resistance influence trend are highly consistent, it may indicate that the performance fluctuations are caused by changes in slurry characteristics.

[0043] The cloud computing-based centrifugal pump performance evaluation method proposed in this application can effectively distinguish between performance fluctuations caused by changes in slurry characteristics and performance degradation caused by physical wear of the pump body by introducing the key parameter of fluid resistance influence and combining its changing trend.

[0044] Specifically, this method first collects real-time operating data of the centrifugal pump, including inlet pressure, outlet pressure, liquid flow rate, motor input current, and motor input voltage. This data is then transmitted to a cloud platform for processing.

[0045] Next, based on this operational data, the actual motor input power, actual head, and actual efficiency of the centrifugal pump are precisely calculated. These actual performance parameters reflect the pump's true operating status under current conditions.

[0046] Simultaneously, using preset clean water baseline performance data, the theoretical clean water head and theoretical clean water efficiency corresponding to the current liquid flow rate are obtained, and the theoretical clean water motor input power is calculated. These theoretical values ​​represent the pump's performance under ideal clean water conditions.

[0047] Subsequently, the difference between the actual motor input power and the theoretical clean water motor input power was compared to calculate the influence of fluid resistance. This influence directly quantifies the impact of the difference in fluid characteristics between the actual conveyed medium (such as slurry) and the clean water medium on pump power consumption.

[0048] Based on this, the theoretical clean water head and theoretical clean water efficiency are corrected using the calculated influence of fluid resistance, resulting in target theoretical clean water head and target theoretical clean water efficiency. This correction step makes the theoretical performance benchmark closer to actual operating conditions, thereby improving the accuracy of subsequent performance evaluations.

[0049] Finally, the actual head and efficiency are compared with the corrected target theoretical head and efficiency for clean water to obtain the performance deviation. More importantly, a comprehensive judgment is made by combining the performance deviation with the changing trends of the fluid resistance effect. For example, when the performance deviation continues to increase, while the fluid resistance effect remains stable or shows fluctuations inconsistent with the performance deviation, the system can accurately diagnose physical wear of the pump body. Conversely, if the performance deviation and the fluid resistance effect show highly consistent trends, it indicates that the performance fluctuation is mainly caused by changes in the slurry characteristics.

[0050] Through the aforementioned collaborative efforts, the method of this application overcomes the limitations of existing technologies that cannot distinguish between performance degradation caused by changes in slurry characteristics and pump wear. Traditional systems, lacking an understanding of the dynamic changes in the fluid medium, attribute all performance degradation to physical damage to the pump, leading to false alarms and unnecessary maintenance. This application, by introducing the influence of fluid resistance and performing dynamic analysis, can accurately identify the root cause of performance degradation, thereby avoiding misjudgments and improving the accuracy and efficiency of maintenance decisions. For example, when an increase in slurry viscosity causes a temporary performance decline, this application can identify that it is caused by an increase in the influence of fluid resistance, rather than pump wear, thus avoiding unnecessary downtime. When slow wear occurs in the pump body, even if the slurry characteristics temporarily mask some performance degradation, this application can promptly detect and warn of actual physical damage by observing the stability of the influence of fluid resistance and the continuous deterioration trend of performance deviation, ensuring production continuity and equipment reliability.

[0051] In some embodiments, the preset clean water baseline performance data is stored in the cloud in the form of discrete points, including the theoretical clean water head and theoretical clean water efficiency corresponding to different liquid flow rates. The theoretical clean water head and theoretical clean water efficiency are obtained by table lookup or linear interpolation.

[0052] Specifically, the preset clean water baseline performance data refers to the ideal performance parameters of the centrifugal pump when operating under standard clean water medium, typically obtained through experimental testing or design calculations. This data is stored as discrete points, rather than continuous curves, to optimize storage space and data management efficiency. Each discrete point corresponds to a specific liquid flow rate, and records the theoretical clean water head and theoretical clean water efficiency at that flow rate. Storing this data in the cloud enables centralized data management, remote access, and distributed processing, thereby supporting performance evaluation of a large number of centrifugal pumps.

[0053] Furthermore, when it is necessary to obtain the theoretical clean water head and theoretical clean water efficiency corresponding to the current liquid flow rate, either a lookup table or linear interpolation can be used. A lookup table involves directly searching the stored discrete point data for a record that precisely matches the current liquid flow rate and directly obtaining its corresponding theoretical clean water head and theoretical clean water efficiency. When the current liquid flow rate does not precisely match any discrete point, linear interpolation is used. Linear interpolation involves selecting two discrete points adjacent to the current liquid flow rate and calculating the theoretical clean water head and theoretical clean water efficiency under the current liquid flow rate through a linear relationship based on their respective theoretical clean water head and theoretical clean water efficiency.

[0054] This application's solution stores clean water benchmark performance data in discrete point form in the cloud and retrieves it using table lookup or linear interpolation, ensuring the accuracy and accessibility of the benchmark data. Therefore, when evaluating centrifugal pump performance, the required theoretical clean water head and theoretical clean water efficiency can be obtained quickly and accurately, providing a reliable data foundation for subsequent performance calculations and diagnostics. At the same time, the cloud storage model also enhances data management flexibility and system scalability.

[0055] The above technical solutions effectively improve the efficiency and accuracy of acquiring water quality benchmark performance data, avoiding the need for complex curve fitting or large amounts of local storage that may exist in traditional methods. Furthermore, storing data in the cloud allows evaluation systems in different geographical locations to share and utilize the latest benchmark data, reducing data maintenance costs and enhancing the robustness and scalability of the entire performance evaluation system.

[0056] In some embodiments, the step of calculating the influence of fluid resistance based on the actual motor input power and the theoretical clean water motor input power includes:

[0057] Monitor the operating status of the centrifugal pump;

[0058] Based on the operating status, determine whether the centrifugal pump is in a transient operating condition;

[0059] When the centrifugal pump is in transient operation, the calculation of the fluid resistance effect is suspended or the data under transient operation is marked as invalid.

[0060] When the centrifugal pump is running in steady state, the difference between the actual motor input power and the theoretical clean water motor input power is used as the fluid resistance influence quantity, and the fluid resistance influence quantity is averaged.

[0061] Determining whether a centrifugal pump is in a transient operating condition can be achieved by setting threshold values ​​for the rate of change of various operating parameters. For example, when the liquid flow rate, motor input current, or motor input voltage changes by more than a preset threshold within a short period of time, the pump can be identified as being in a transient operating condition. The purpose is to accurately identify the unstable phase of pump operation.

[0062] When a centrifugal pump is in transient operation, suspending the calculation of fluid resistance influence means that the calculation of fluid resistance influence is not performed during this period to avoid interference from transient fluctuations in the calculation results. Alternatively, data under transient conditions is marked as invalid, meaning that although the data is collected, it is not used for the calculation of fluid resistance influence, thus filtering out unreliable information at the data processing level. The purpose is to prevent abnormal data under transient conditions from contaminating the calculation results of fluid resistance influence, ensuring data purity. When the centrifugal pump is in steady-state operation, the difference between the actual motor input power and the theoretical clean water motor input power is used as the fluid resistance influence, and this influence is averaged. Specifically, steady-state operation means that the various operating parameters of the pump remain relatively stable within a certain time window, with fluctuations within an acceptable range. Under this stable condition, the difference between the actual motor input power and the theoretical clean water motor input power can more accurately reflect the additional energy consumption caused by fluid resistance. Averaging this difference, such as using moving average or exponential average methods, can further eliminate random noise that may exist in steady-state operation, improving the smoothness and reliability of the fluid resistance influence calculation. The aim is to obtain a stable and representative value for the influence of fluid resistance.

[0063] Through the above technical solution, this application can significantly improve the accuracy and robustness of the calculation of fluid resistance influence. By intelligently identifying and processing data under transient operating conditions, calculation errors introduced by instantaneous fluctuations and unsteady-state effects are avoided, thus ensuring that the fluid resistance influence obtained under steady-state conditions is more realistic and representative. This accurate fluid resistance influence allows for a more accurate identification of performance fluctuations caused by changes in fluid characteristics when comparing actual head and efficiency with target theoretical clean water head and efficiency, and a clearer identification of performance degradation caused by pump body physical wear, greatly improving the reliability and diagnostic accuracy of centrifugal pump performance evaluation.

[0064] For example, suppose a centrifugal pump is used in a chemical plant to transport slurry containing solid particles. During the pump startup phase, the motor input current and liquid flow rate rise rapidly in a short period of time, at which point the system determines that the pump is in a transient operating condition. During this transient period, the calculation of the fluid resistance influence is suspended, or the data collected during this period is marked as invalid and not included in subsequent calculations. After the pump has been running for about 5 minutes, the various operating parameters tend to stabilize, and the liquid flow rate, inlet pressure, outlet pressure, etc., all fluctuate within the preset steady-state range. At this point, the system determines that the pump has entered steady-state operation. During steady-state operation, the system continuously acquires the actual motor input power and the theoretical clean water motor input power, and calculates the difference between the two. For example, the difference is calculated every 10 seconds, and then all differences within the past minute are processed by a moving average to obtain a smooth fluid resistance influence. In this way, the interference of the instantaneous impact during the startup phase on the calculation of the fluid resistance influence can be effectively avoided, ensuring that the obtained fluid resistance influence can accurately reflect the true impact of slurry characteristics or pump wear under steady-state operation.

[0065] In some embodiments, this application further proposes the steps of comparing the actual head and actual efficiency with the target theoretical clean water head and target theoretical clean water efficiency to obtain the comparison results, and distinguishing between performance fluctuations caused by changes in slurry characteristics and performance degradation caused by physical wear of the pump body based on the comparison results and the changing trend of the fluid resistance influence.

[0066] The performance deviation between the actual head and actual efficiency and the target theoretical clean water head and target theoretical clean water efficiency is obtained.

[0067] Based on the changing trends of the performance deviation and the fluid resistance influence, it is determined whether the performance deviation continues to exceed the preset range and whether the fluid resistance influence is in a low fluctuation state.

[0068] When the performance deviation continues to exceed the preset range and the fluid resistance influence is in a low fluctuation state, it is determined to be physical wear of the pump body;

[0069] When the performance deviation is highly consistent with the trend of the change in the fluid resistance effect, it is determined to be a performance fluctuation caused by the change in the slurry characteristics;

[0070] When the performance deviation continues to exceed the preset range, and the fluid resistance influence shows a fluctuation trend inconsistent with the performance deviation, the long-term deterioration rate of the performance deviation is analyzed, and based on the long-term deterioration rate, it is determined whether the pump body is experiencing accelerated wear.

[0071] Specifically, performance deviation refers to the difference between the actual head and actual efficiency of a centrifugal pump and the target theoretical clean water head and target theoretical clean water efficiency. This deviation can be quantitatively reflected as the gap between the performance of the centrifugal pump under current operating conditions and the ideal benchmark. For example, performance deviation can be obtained by calculating the percentage difference between the actual head and the target theoretical clean water head, and the percentage difference between the actual efficiency and the target theoretical clean water efficiency.

[0072] The determination of whether performance deviations consistently exceed the preset range aims to identify whether performance degradation has reached a level requiring attention, and whether this degradation is not accidental or transient. The preset range can be set based on pump type, operating conditions, historical data, and experience. Its purpose is to filter out normal operating fluctuations and focus on abnormal performance. When the influence of fluid resistance is in a low-fluctuation state, it means that the characteristics of the fluid medium (such as concentration and viscosity) are relatively stable. In this case, changes in performance deviation are more likely to be attributed to the physical state of the pump itself.

[0073] In practical applications, when performance deviations consistently exceed the preset range and the influence of fluid resistance remains low, this strongly indicates that the pump body may be experiencing physical wear. This is because, under stable medium characteristics, a continuous decline in performance is usually due to wear of internal pump components (such as impellers and pump casings), leading to a decrease in hydraulic performance.

[0074] Furthermore, when the performance deviation and the influence of fluid resistance show a highly consistent trend—for example, both rise or fall simultaneously with similar amplitudes—this indicates that the performance fluctuation is primarily caused by changes in slurry characteristics. Changes in slurry characteristics directly affect fluid resistance, which in turn affects the pump's actual head and efficiency.

[0075] Furthermore, when performance deviations consistently exceed the preset range, but the fluid resistance effect exhibits a fluctuating trend inconsistent with the performance deviation—for example, performance continuously declines while the fluid resistance effect remains stable or fluctuates only slightly—further analysis of the long-term deterioration rate of the performance deviation is necessary. Analyzing the long-term deterioration rate can reveal the speed and severity of performance degradation, thereby determining whether accelerated wear exists in the pump body. Accelerated wear usually indicates that the internal wear mechanism of the pump body is intensifying, requiring more urgent intervention.

[0076] This application's solution, by introducing a judgment of "continuously exceeding the preset range" in performance deviation, a judgment of "low fluctuation state" in the influence of fluid resistance, and an analysis of "high consistency in the changing trends" of the two, can more precisely identify the root cause of centrifugal pump performance degradation. Specifically, by monitoring whether the performance deviation continuously exceeds the preset range, short-term, accidental performance fluctuations can be effectively ruled out, focusing on long-term, systemic performance problems. Simultaneously, by combining the fluctuation state of the influence of fluid resistance, performance fluctuations caused by changes in media characteristics can be effectively decoupled from performance degradation caused by the pump's own physical wear. When the influence of fluid resistance is in a low fluctuation state, the continuous performance decline more directly points to pump wear; while when the trends of the two are highly consistent, it clearly indicates that media characteristics are the dominant factor. Furthermore, for complex situations, by introducing long-term deterioration rate analysis of performance deviation, it is possible to further identify whether accelerated wear exists, thus providing a basis for earlier warning and intervention. This multi-dimensional, trend-based judgment mechanism significantly improves the accuracy and reliability of diagnosis.

[0077] Through the above technical solution, this application overcomes the ambiguity of traditional methods in distinguishing the causes of centrifugal pump performance fluctuations, significantly improving the accuracy and reliability of performance evaluation. Specifically, by introducing a comprehensive judgment on the persistence of performance deviation, the fluctuation state of fluid resistance influence, and the consistency of their trends, it can effectively avoid misjudging short-term performance fluctuations caused by changes in slurry characteristics as physical wear of the pump body, or vice versa. Especially when performance deviations continuously exceed the preset range but the fluctuation of fluid resistance influence is inconsistent, by analyzing the long-term deterioration rate of performance deviation, it is possible to promptly detect and warn of potential accelerated wear of the pump body, thereby providing a more accurate basis for predictive maintenance, extending equipment service life, and reducing operating costs.

[0078] In some preferred embodiments, a specific example is given below. Suppose a centrifugal pump is conveying a slurry containing solid particles, and its operating data is continuously collected and uploaded to the cloud for evaluation. Over a certain period of time, the system detects a continuous performance deviation between the actual head and actual efficiency and the target theoretical clean water head and target theoretical clean water efficiency, and this deviation exceeds a preset range of 5%.

[0079] At this point, the system will further analyze the changing trend of the influence of fluid resistance.

[0080] Scenario 1: If the fluid resistance effect also shows a decreasing trend highly consistent with the performance deviation within the same time period, for example, both decreasing by about 5%, the system will determine that the performance decline is caused by performance fluctuations due to changes in slurry characteristics such as increased slurry concentration or viscosity. In this case, it may be necessary to adjust the slurry formulation or pump operating parameters, rather than immediately performing pump overhaul.

[0081] Scenario 2: If the fluid resistance effect is found to remain low-fluctuation while the performance deviation continues to exceed the preset range, meaning its variation is much smaller than the performance deviation (e.g., the fluid resistance effect fluctuates within 1%), the system will determine that the performance degradation is caused by physical wear of the pump body. This may indicate that components such as the impeller or pump casing are worn, requiring a shutdown for inspection and maintenance.

[0082] Scenario 3: If the performance deviation continues to exceed the preset range, but the fluid resistance effect shows a fluctuating trend inconsistent with the performance deviation—for example, the performance deviation continues to worsen while the fluid resistance effect fluctuates without clear consistency—the system will further analyze the long-term deterioration rate of the performance deviation. If the calculated long-term deterioration rate continues to exceed the preset accelerated wear threshold (e.g., performance decreases by more than 1% per month), the system will issue an accelerated wear warning, indicating potential serious pump damage and requiring immediate detailed inspection to avoid sudden failures and greater losses.

[0083] Through the specific judgment logic described above, the evaluation method of this application can provide more accurate and timely guidance for the operation and maintenance of centrifugal pumps.

[0084] In some embodiments, this application further proposes the following steps: when the performance deviation continuously exceeds a preset range and the fluid resistance influence shows a fluctuation trend inconsistent with the performance deviation, analyzing the long-term deterioration rate of the performance deviation, and determining whether the pump body experiences accelerated wear based on the long-term deterioration rate:

[0085] Short-time-scale smoothing is performed on the performance deviation data stream to obtain short-term smoothed performance deviation.

[0086] Long-term smoothing performance deviation is obtained by performing long-term smoothing on the short-term smoothing performance deviation data stream.

[0087] On the long-term smoothing performance deviation data stream, sampling is performed at fixed time intervals to obtain long-term smoothing performance deviation sampling points;

[0088] Calculate the difference between adjacent long-term smooth performance deviation sampling points and divide it by the fixed time interval to obtain the long-term deterioration rate of the performance deviation;

[0089] The long-term deterioration rate is compared with a preset accelerated wear threshold. When the long-term deterioration rate continues to exceed the accelerated wear threshold, it is determined that the pump body is experiencing accelerated wear.

[0090] Specifically, the performance deviation data stream refers to the time-varying sequence data showing the difference between the actual head and efficiency of a centrifugal pump and the target theoretical head and efficiency for clean water. Short-timescale smoothing of this performance deviation data stream aims to filter out high-frequency noise and transient disturbances, resulting in a smoothed performance deviation and thus a clearer data trend. For example, moving averages, exponential smoothing, or Savitzky-Golay filters can be used for this process.

[0091] Furthermore, long-term smoothing is applied to the short-term smoothed performance deviation data stream to extract the long-term trend of the performance deviation, eliminate the impact of short- and medium-term fluctuations, and obtain a long-term smoothed performance deviation. The window or parameter of this long-term smoothing is usually larger than that of the short-term smoothing to better capture long-term changes.

[0092] After acquiring the long-term smoothed performance deviation data stream, sampling at fixed time intervals yields a series of discrete long-term smoothed performance deviation sampling points. These sampling points represent long-term performance trend values ​​at specific time points.

[0093] Subsequently, by calculating the difference between adjacent long-term smoothed performance deviation sampling points and dividing it by the fixed time interval, the long-term deterioration rate of the performance deviation can be obtained. This rate reflects the average rate of decline in pump performance over time.

[0094] Finally, the calculated long-term deterioration rate is compared with a preset accelerated wear threshold. This accelerated wear threshold is a critical value set based on experience, historical data, or expert knowledge, used to determine whether pump body wear has entered an accelerated phase. When the long-term deterioration rate consistently exceeds this threshold, it can be determined that the pump body is experiencing accelerated wear, indicating the need for further inspection or maintenance.

[0095] This application's solution effectively addresses the interference of noise and short-term fluctuations in the original performance deviation data on the calculation of long-term degradation rates by introducing short-term and long-term smoothing processes. Short-term smoothing initially filters out high-frequency noise, resulting in a more stable trend in the data within a shorter time window. Building upon this, long-term smoothing further extracts the long-term evolution trend of the performance deviation, thus avoiding misleading assessments of long-term degradation rates by short-to-medium-term operating condition changes or measurement errors. This multi-scale smoothing process allows the calculated long-term degradation rate to more accurately reflect the actual physical wear process of the pump body, rather than surface fluctuations influenced by instantaneous factors. By comparing this stable and accurate long-term degradation rate with a preset accelerated wear threshold, it is possible to reliably identify whether the pump body has entered the accelerated wear stage, providing a solid data foundation for timely intervention.

[0096] Through the above technical solution, this application effectively overcomes the limitations of traditional methods in assessing accelerated pump wear, which are susceptible to data noise and short-term fluctuations. By performing multi-scale smoothing on the performance deviation data stream, the long-term deterioration trend of pump performance can be extracted more accurately, making the calculated long-term deterioration rate more stable and reliable. Therefore, when determining whether accelerated wear exists in the pump body, the accuracy and timeliness of diagnosis can be significantly improved, avoiding equipment failures or production losses due to misjudgment or omission, and providing a more scientific and effective basis for predictive maintenance of centrifugal pumps.

[0097] In some preferred embodiments, it is assumed that the performance deviation data stream of a centrifugal pump is continuously monitored during operation. To accurately determine whether accelerated wear of the pump body exists, the performance deviation data stream is first smoothed on a short time scale. For example, a 5-point moving average filter can be used to average every 5 consecutive performance deviation data points to eliminate high-frequency noise and obtain a short-term smoothed performance deviation. Subsequently, to further extract the long-term trend, the short-term smoothed performance deviation data stream is smoothed on a long time scale, for example, using a 50-point exponential smoothing filter to smooth the data more significantly and obtain a long-term smoothed performance deviation.

[0098] Next, sampling is performed on the long-term smoothed performance deviation data stream at fixed time intervals of one hour to obtain a series of long-term smoothed performance deviation sampling points. For example, sampling point values ​​P_T and P_{T+1} are obtained at hour T and hour T+1, respectively. Then, the difference between adjacent sampling points (P_{T+1} - P_T) is calculated and divided by the fixed time interval (1 hour) to obtain the long-term deterioration rate of performance deviation within that hour.

[0099] Finally, the calculated long-term degradation rate is compared with a preset accelerated wear threshold (e.g., a performance decrease of 0.05% per hour). If the long-term degradation rate consistently exceeds 0.05% for multiple consecutive sampling periods (e.g., 24 consecutive hours), the system will determine that the pump body is experiencing accelerated wear and trigger corresponding warnings or maintenance recommendations.

[0100] In some embodiments, when the performance deviation continuously exceeds a preset range and the fluid resistance effect exhibits a fluctuation trend inconsistent with the performance deviation, the step of analyzing the long-term deterioration rate of the performance deviation and determining whether the pump body experiences accelerated wear based on the long-term deterioration rate includes:

[0101] Short-time-scale smoothing is performed on the performance deviation data stream to obtain short-term smoothed performance deviation.

[0102] Long-term smoothing performance deviation is obtained by performing long-term smoothing on the short-term smoothing performance deviation data stream.

[0103] Periodic features are extracted from long-term smoothing performance deviation data streams to identify periodic fluctuation components present in the data streams.

[0104] The non-periodic trend performance deviation is obtained by subtracting the periodic fluctuation component from the long-term smoothing performance deviation data stream.

[0105] On the non-periodic trend performance deviation data stream, sampling is performed at fixed time intervals to obtain non-periodic trend performance deviation sampling points;

[0106] Calculate the difference between adjacent sampling points of the non-periodic trend performance deviation and divide it by the fixed time interval to obtain the long-term deterioration rate;

[0107] The long-term deterioration rate is compared with a preset accelerated wear threshold. When the long-term deterioration rate continues to exceed the accelerated wear threshold, it is determined that the pump body is experiencing accelerated wear.

[0108] Specifically, short-timescale smoothing is performed on the performance deviation data stream to eliminate instantaneous noise and high-frequency random fluctuations, thereby obtaining short-term smoothed performance deviations. This processing can be achieved using methods such as moving averages and exponential smoothing, with the aim of improving the local stability of the data. Further, long-timescale smoothing is performed on the short-term smoothed performance deviation data stream to further filter out medium- and high-frequency fluctuations, obtaining long-term smoothed performance deviations. This step helps reveal more macroscopic trend changes in the data, providing a more stable foundation for subsequent periodic feature extraction.

[0109] The key to this solution is extracting periodic features from the long-term smoothed performance deviation data stream to identify the periodic fluctuation components. This step aims to accurately identify and quantify periodic changes in the data caused by non-wear factors, such as diurnal temperature variations, seasonal load changes, or periodic maintenance operations. In practical applications, subtracting the periodic fluctuation components from the long-term smoothed performance deviation data stream effectively isolates the purely non-periodic trend performance deviation. This operation ensures that subsequent deterioration rate calculations more accurately reflect the physical wear trend of the pump body itself, avoiding interference from periodic factors. Sampling at fixed time intervals on the non-periodic trend performance deviation data stream captures its changes at discrete time points. The difference between adjacent non-periodic trend performance deviation sampling points is then calculated and divided by the fixed time interval to obtain the long-term deterioration rate of the performance deviation. This rate directly quantifies the degree to which the pump body's performance deteriorates over time. Finally, the long-term deterioration rate is compared with a preset accelerated wear threshold. When the long-term deterioration rate consistently exceeds the accelerated wear threshold, it can be determined that the pump body is experiencing accelerated wear. The accelerated wear threshold is preset based on empirical data or factors such as the pump body's design life, and is used to define whether the performance degradation has reached a level that requires attention.

[0110] This application's solution addresses the problem of traditional methods' difficulty in accurately assessing the long-term deterioration rate of pump bodies when periodic fluctuations exist by introducing periodic feature extraction and component stripping steps. Specifically, the original performance deviation data stream is first subjected to multi-scale smoothing to eliminate noise and short-term fluctuations, resulting in a long-term smoothed performance deviation. Based on this, periodic feature extraction is performed on the long-term smoothed performance deviation data stream to identify and quantify periodic fluctuation components caused by non-wear factors (such as operating cycles and environmental changes). Subsequently, these periodic fluctuation components are subtracted from the long-term smoothed performance deviation data stream, resulting in a purer aperiodic trend performance deviation data stream. This aperiodic trend performance deviation data stream more realistically reflects the performance degradation trend of the pump body due to physical wear, eliminating periodic interference. By sampling and differentially calculating this trend data with periodic components removed, a more accurate long-term deterioration rate can be obtained, which can then be compared with an accelerated wear threshold to accurately determine whether the pump body experiences accelerated wear.

[0111] In some preferred embodiments, it is assumed that during the operation of a centrifugal pump, its performance deviation data stream, in addition to the slow decreasing trend caused by pump wear, also exhibits significant daily periodic fluctuations (e.g., high daytime load and low nighttime load causing periodic changes in performance deviation within a day) and weekly periodic fluctuations (e.g., weekend maintenance shutdowns leading to a brief performance recovery at the beginning of the week). First, the original performance deviation data stream is smoothed on a short timescale, for example, using a 3-hour moving average, to filter out instantaneous noise, resulting in a short-term smoothed performance deviation. Next, the short-term smoothed performance deviation data stream is smoothed on a long timescale, for example, using a 24-hour moving average, to obtain a long-term smoothed performance deviation. Subsequently, periodic features are extracted from the long-term smoothed performance deviation data stream. For example, methods such as Empirical Mode Decomposition (EMD) or wavelet analysis can be used to decompose it into multiple Intrinsic Mode Functions (IMFs) and identify the IMFs corresponding to the daily and weekly cycles. By performing energy analysis on these IMFs, their amplitude and phase can be determined, thereby reconstructing the overall periodic fluctuation components. Next, the reconstructed periodic fluctuation component is subtracted from the long-term smoothed performance deviation data stream to obtain a non-periodic trend performance deviation free from periodic effects. This trend data will more clearly show the performance degradation caused by pump wear. Finally, the non-periodic trend performance deviation data stream is sampled at fixed time intervals, such as once a week. The difference between adjacent sampling points is calculated and divided by the time interval to obtain the long-term degradation rate. This rate is compared with a preset accelerated wear threshold. If the rate consistently exceeds the accelerated wear threshold, it is accurately determined that the pump body is experiencing accelerated wear, without being disturbed by daily or weekly periodic fluctuations.

[0112] In some embodiments, the step of extracting periodic features from the long-term smoothing performance deviation data stream and identifying periodic fluctuation components present in the long-term smoothing performance deviation data stream may include the following:

[0113] Multi-scale spectral decomposition is performed on the long-term smoothing performance deviation data stream, decomposing it into multiple independent frequency components;

[0114] Energy analysis is performed on each decomposed frequency component to identify frequency components with significant energy peaks;

[0115] Based on the identified frequency components with significant energy peaks, determine the period and amplitude of their corresponding periodic fluctuation components;

[0116] Based on the determined period and amplitude, multiple periodic fluctuation components are superimposed and reconstructed to obtain the periodic fluctuation components present in the long-term smoothing performance deviation data stream.

[0117] Specifically, multi-scale spectral decomposition refers to decomposing a complex long-term smoothing performance deviation data stream into a series of simple signal components with different frequencies and amplitudes. For example, methods such as Empirical Mode Decomposition (EMD), wavelet transform, or Fourier transform can be used. The aim is to separate various periodic or quasi-periodic patterns in the original signal for independent subsequent analysis. Each decomposed frequency component can be understood as a fluctuation pattern of the original signal within a specific frequency range; ideally, these components are independent of each other.

[0118] Furthermore, energy analysis is performed on each decomposed frequency component to quantify its importance. An energy distribution curve is obtained by calculating the energy distribution of each frequency component within a preset time window. Energy peaks are identified within this curve; these peaks typically correspond to periodic fluctuations with significant energy in the signal. By analyzing the peak width and peak height, and combining this with a preset noise characteristic threshold, it can be determined whether the energy peaks represent true periodic fluctuations, thereby identifying frequency components with significant energy peaks.

[0119] Therefore, based on the identified frequency components with significant energy peaks, the period and amplitude of their corresponding periodic fluctuation components can be determined. For example, for the spectrum obtained through Fourier transform, the frequency corresponding to the energy peak is the frequency of the periodic component, and the period can then be calculated; the height of the peak reflects the amplitude of the periodic component. The aim is to accurately quantify the characteristics of each major periodic fluctuation component.

[0120] Finally, based on the determined period and amplitude, the identified multiple periodic fluctuation components are superimposed and reconstructed to obtain the periodic fluctuation components present in the long-term smoothing performance deviation data stream. This means that all identified independent frequency components with significant periodic characteristics are linearly superimposed according to their respective periods and amplitudes to synthesize a comprehensive signal representing all periodic fluctuations in the original signal.

[0121] Through the above technical solution, this application can accurately extract the periodic fluctuation component from a long-term smooth performance deviation data stream. Compared with solutions that do not perform fine periodic feature extraction, this application can more effectively distinguish between performance fluctuations caused by periodic factors and performance degradation caused by physical wear of the pump body. This significantly improves the calculation accuracy of non-periodic trend performance deviations, thereby making the judgment on whether the pump body is experiencing accelerated wear more accurate and reliable, avoiding misjudgments or omissions caused by periodic interference, and providing a more solid data foundation for predictive maintenance of centrifugal pumps.

[0122] In some embodiments, the step of performing multi-scale spectral decomposition on the long-term smoothing performance deviation data stream to decompose it into multiple independent frequency components may include:

[0123] Local extrema are identified in data streams with long-term smoothing performance deviations.

[0124] Construct the upper and lower envelopes of the long-term smoothing performance deviation data stream based on the local extreme points;

[0125] The intrinsic mode functions are extracted iteratively using the upper and lower envelopes, and each intrinsic mode function is used as a frequency component of the long-term smoothing performance deviation data stream, until the residual component of the long-term smoothing performance deviation data stream is a monotonic function or the number of extreme points of the residual component is less than a preset value.

[0126] Specifically, the above steps aim to decompose a complex long-term smoothed performance deviation data stream into a series of physically meaningful intrinsic mode functions using an adaptive data-driven approach. Identifying local extrema refers to determining, within the long-term smoothed performance deviation data stream, the points that represent the maximum or minimum value relative to their neighbors. These local extrema are crucial for constructing the data envelope. Furthermore, based on the identified local extrema, upper and lower envelopes of the long-term smoothed performance deviation data stream can be constructed. The upper envelope is obtained by connecting all local maxima points and performing smooth interpolation, while the lower envelope is obtained by connecting all local minima points and performing smooth interpolation.

[0127] Based on this, intrinsic mode functions (IMFs) are iteratively extracted using the upper and lower envelopes. An IMF is defined as a function satisfying two conditions: first, the number of extrema and the number of zero-crossings must be equal or differ by at most one throughout the entire data sequence; second, at any point, the average of the upper envelope defined by local maxima and the lower envelope defined by local minima is zero. The iterative extraction process, often referred to as a "screening process," progressively separates the IMFs by subtracting the envelope mean formed by the averages of the upper and lower envelopes from the original signal. This process is repeated until the extracted components satisfy the IMF conditions. Each successfully extracted IMF is considered a frequency component of the long-term smoothing performance deviation data stream, representing a fluctuation pattern at a specific scale in the data. This iterative process continues until the residual components of the long-term smoothing performance deviation data stream are monotonic, i.e., no longer contain any oscillations, or the number of extrema of the residual components is less than a preset value, indicating that no more meaningful IMFs can be extracted.

[0128] The above technical solution enables a refined decomposition of the centrifugal pump performance deviation data stream, overcoming the limitations of traditional Fourier transform and other methods in processing nonlinear and non-stationary signals. This adaptive decomposition capability allows the system to more accurately identify performance fluctuations caused by different factors (such as changes in slurry characteristics and physical wear of the pump body), especially in the presence of complex multi-scale oscillations. This provides more reliable and accurate basic data for subsequent periodic feature extraction and accelerated wear assessment, significantly improving the accuracy and robustness of centrifugal pump performance evaluation.

[0129] In some embodiments, in the step of performing energy analysis on each decomposed frequency component and identifying frequency components with significant energy peaks, this application further proposes the following specific implementation methods.

[0130] For each decomposed frequency component, calculate its energy distribution within a preset time window to obtain the energy distribution curve;

[0131] Identify the energy peak point in the energy distribution curve;

[0132] For each energy peak point, calculate its peak width and peak height;

[0133] Based on the peak width and peak height, and combined with a preset noise characteristic threshold, it is determined whether the energy peak point satisfies the true periodic fluctuation.

[0134] When the energy peak point satisfies the characteristics of true periodic fluctuation, the frequency component is identified as a frequency component with a significant energy peak.

[0135] The process of calculating the energy distribution of each decomposed frequency component within a preset time window to obtain an energy distribution curve involves quantifying the energy of each independent frequency component obtained through multi-scale spectral decomposition over a specific time period. The preset time window can be a fixed-length sliding window used to dynamically capture energy changes in the frequency components. This generates a curve reflecting the energy change of the frequency component over time, i.e., the energy distribution curve. Furthermore, identifying energy peak points in the energy distribution curve involves detecting local maxima in the curve using an algorithm. These peak points indicate a significant increase in the energy of the frequency component at a specific moment or time period.

[0136] Specifically, for each energy peak point, its peak width and peak height are calculated. The peak height refers to the maximum energy value at the energy peak point, while the peak width can be understood as the duration or frequency range during which the energy peak point remains above a certain energy threshold (e.g., 50% of the peak height). These parameters are used to quantify the intensity and duration of the energy peak. As a preferred embodiment, based on the peak width and peak height, combined with a preset noise characteristic threshold, it is determined whether the energy peak point satisfies true periodic fluctuation. The noise characteristic threshold may include minimum peak height, minimum peak width, or signal-to-noise ratio, etc., and its purpose is to distinguish between periodic fluctuations caused by real physical processes and spurious peaks caused by random noise or transient interference. When the energy peak point satisfies the characteristics of true periodic fluctuation, that is, when both its peak height and peak width exceed the preset threshold, and it may also satisfy other noise suppression conditions, then the frequency component is identified as a frequency component with a significant energy peak.

[0137] Through the above technical solution, this application can more accurately identify the true periodic fluctuation component from complex long-term smoothed performance deviation data streams. Compared with simply performing spectral decomposition, this solution, by introducing the calculation of energy distribution curves, peak width, and peak height, and combining this with noise characteristic thresholds for judgment, effectively avoids misjudging noise or non-periodic transient events as periodic fluctuations, thereby improving the accuracy of periodic feature extraction. This provides a solid foundation for subsequently subtracting periodic fluctuation components from the long-term smoothed performance deviation data stream to obtain a purer non-periodic trend performance deviation, thus making the judgment of accelerated pump wear more reliable and accurate.

[0138] like Figure 2 As shown, an exemplary cloud-based centrifugal pump performance evaluation system is illustrated. A specific embodiment of this application discloses a cloud-based centrifugal pump performance evaluation system 100, comprising:

[0139] The power calculation module 10 is used to calculate the actual motor input power, actual head and actual efficiency based on the operating data of the centrifugal pump. The operating data includes inlet pressure, outlet pressure, liquid flow rate, motor input current and motor input voltage.

[0140] The benchmark establishment module 20 is used to obtain the theoretical clean water head and theoretical clean water efficiency corresponding to the liquid flow rate based on the preset clean water benchmark performance data, and to calculate the theoretical clean water motor input power based on the theoretical clean water head and the theoretical clean water efficiency.

[0141] The resistance influence calculation module 30 is used to calculate the fluid resistance influence based on the actual motor input power and the theoretical clean water motor input power.

[0142] The standard correction module 40 is used to correct the theoretical clean water head and theoretical clean water efficiency based on the fluid resistance influence amount, and generate the target theoretical clean water head and target theoretical clean water efficiency.

[0143] The performance diagnostic module 50 is used to compare the actual head and actual efficiency with the target theoretical clean water head and target theoretical clean water efficiency to obtain the comparison results, and to distinguish between performance fluctuations caused by changes in slurry characteristics and performance degradation caused by physical wear of the pump body based on the comparison results and the changing trend of the fluid resistance influence.

[0144] The cloud-based centrifugal pump performance evaluation system proposed in this application introduces a resistance influence calculation module and a standard correction module, combined with a performance diagnosis module to analyze the changing trends of fluid resistance influence, enabling accurate identification of the root cause of performance degradation. For example, when increased slurry viscosity leads to a temporary performance decrease, the system can identify that this is caused by increased fluid resistance influence, rather than pump wear, thus avoiding unnecessary downtime. When slow wear occurs in the pump body, even if slurry characteristics temporarily mask some performance degradation, the system can still detect and warn of actual physical damage in a timely manner by observing the stability of fluid resistance influence and the continuous deterioration trend of performance deviation, ensuring production continuity and equipment reliability. This modular system design makes the responsibilities of each functional unit clear, easy to deploy, maintain, and upgrade, significantly improving the accuracy and efficiency of centrifugal pump performance evaluation.

[0145] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A cloud computing-based method for evaluating the performance of centrifugal pumps, characterized in that, include: Based on the centrifugal pump's operating data, calculate the actual motor input power, actual head, and actual efficiency. The operating data includes the centrifugal pump's inlet pressure, outlet pressure, liquid flow rate, motor input current, and motor input voltage. Based on the preset clean water benchmark performance data, obtain the theoretical clean water head and theoretical clean water efficiency corresponding to the liquid flow rate, and calculate the theoretical clean water motor input power based on the theoretical clean water head and the theoretical clean water efficiency. Calculate the influence of fluid resistance based on the actual motor input power and the theoretical clean water motor input power; Based on the fluid resistance influence, the theoretical clean water head and theoretical clean water efficiency are corrected to generate the target theoretical clean water head and target theoretical clean water efficiency. The actual head and actual efficiency are compared with the target theoretical head and target theoretical water efficiency to obtain the comparison results. Based on the comparison results and the changing trend of the fluid resistance influence, the performance fluctuations caused by changes in slurry characteristics and the performance degradation caused by physical wear of the pump body are distinguished.

2. The centrifugal pump performance evaluation method based on cloud computing according to claim 1, characterized in that, The preset clean water benchmark performance data is stored in the cloud in the form of discrete points, including the theoretical clean water head and theoretical clean water efficiency corresponding to different liquid flow rates. The theoretical clean water head and theoretical clean water efficiency are obtained by table lookup or linear interpolation.

3. The centrifugal pump performance evaluation method based on cloud computing according to claim 1, characterized in that, The step of calculating the influence of fluid resistance based on the actual motor input power and the theoretical clean water motor input power includes: Monitor the operating status of the centrifugal pump; Based on the operating status, determine whether the centrifugal pump is in a transient operating condition; When the centrifugal pump is in transient operation, the calculation of the fluid resistance effect is suspended or the data under transient operation is marked as invalid. When the centrifugal pump is running in steady state, the difference between the actual motor input power and the theoretical clean water motor input power is used as the fluid resistance influence quantity, and the fluid resistance influence quantity is averaged.

4. The centrifugal pump performance evaluation method based on cloud computing according to claim 1, characterized in that, The steps of comparing the actual head and actual efficiency with the target theoretical clean water head and target theoretical clean water efficiency to obtain the comparison results, and distinguishing between performance fluctuations caused by changes in slurry characteristics and performance degradation caused by physical wear of the pump body based on the comparison results and the changing trend of the fluid resistance influence, include: The performance deviation between the actual head and actual efficiency and the target theoretical clean water head and target theoretical clean water efficiency is obtained. Based on the changing trends of the performance deviation and the fluid resistance influence, it is determined whether the performance deviation continues to exceed the preset range and whether the fluid resistance influence is in a low fluctuation state. When the performance deviation continues to exceed the preset range and the fluid resistance influence is in a low fluctuation state, it is determined to be physical wear of the pump body; When the performance deviation is highly consistent with the trend of the change in the fluid resistance effect, it is determined to be a performance fluctuation caused by the change in the slurry characteristics; When the performance deviation continues to exceed the preset range, and the fluid resistance influence shows a fluctuation trend inconsistent with the performance deviation, the long-term deterioration rate of the performance deviation is analyzed, and based on the long-term deterioration rate, it is determined whether the pump body is experiencing accelerated wear.

5. The centrifugal pump performance evaluation method based on cloud computing according to claim 4, characterized in that, When the performance deviation continues to exceed a preset range, and the fluid resistance effect exhibits a fluctuation trend inconsistent with the performance deviation, the step of analyzing the long-term deterioration rate of the performance deviation and determining whether the pump body experiences accelerated wear based on the long-term deterioration rate includes: Short-time-scale smoothing is performed on the performance deviation data stream to obtain short-term smoothed performance deviation. Long-term smoothing performance deviation is obtained by performing long-term smoothing on the short-term smoothing performance deviation data stream. On the long-term smoothing performance deviation data stream, sampling is performed at fixed time intervals to obtain long-term smoothing performance deviation sampling points; Calculate the difference between adjacent long-term smooth performance deviation sampling points and divide it by the fixed time interval to obtain the long-term deterioration rate of the performance deviation; The long-term deterioration rate is compared with a preset accelerated wear threshold. When the long-term deterioration rate continues to exceed the accelerated wear threshold, it is determined that the pump body is experiencing accelerated wear.

6. The centrifugal pump performance evaluation method based on cloud computing according to claim 4, characterized in that, When the performance deviation continues to exceed a preset range, and the fluid resistance effect exhibits a fluctuation trend inconsistent with the performance deviation, the step of analyzing the long-term deterioration rate of the performance deviation and determining whether the pump body experiences accelerated wear based on the long-term deterioration rate includes: Short-time-scale smoothing is performed on the performance deviation data stream to obtain short-term smoothed performance deviation. Long-term smoothing performance deviation is obtained by performing long-term smoothing on the short-term smoothing performance deviation data stream. Periodic features are extracted from long-term smoothing performance deviation data streams to identify periodic fluctuation components present in the data streams. The non-periodic trend performance deviation is obtained by subtracting the periodic fluctuation component from the long-term smoothing performance deviation data stream. On the non-periodic trend performance deviation data stream, sampling is performed at fixed time intervals to obtain non-periodic trend performance deviation sampling points; Calculate the difference between adjacent sampling points of the non-periodic trend performance deviation and divide it by the fixed time interval to obtain the long-term deterioration rate; The long-term deterioration rate is compared with a preset accelerated wear threshold. When the long-term deterioration rate continues to exceed the accelerated wear threshold, it is determined that the pump body is experiencing accelerated wear.

7. The centrifugal pump performance evaluation method based on cloud computing according to claim 6, characterized in that, The step of extracting periodic features from the long-term smoothing performance deviation data stream and identifying the periodic fluctuation components present in the long-term smoothing performance deviation data stream includes: Multi-scale spectral decomposition is performed on the long-term smoothing performance deviation data stream, decomposing it into multiple independent frequency components; Energy analysis is performed on each decomposed frequency component to identify frequency components with significant energy peaks; Based on the identified frequency components with significant energy peaks, determine the period and amplitude of their corresponding periodic fluctuation components; Based on the determined period and amplitude, multiple periodic fluctuation components are superimposed and reconstructed to obtain the periodic fluctuation components present in the long-term smoothing performance deviation data stream.

8. The centrifugal pump performance evaluation method based on cloud computing according to claim 7, characterized in that, The step of performing multi-scale spectral decomposition on the long-term smoothing performance deviation data stream, decomposing it into multiple independent frequency components, includes: Local extrema are identified in data streams with long-term smoothing performance deviations. Construct the upper and lower envelopes of the long-term smoothing performance deviation data stream based on the local extreme points; The intrinsic mode functions are extracted iteratively using the upper and lower envelopes, and each intrinsic mode function is used as a frequency component of the long-term smoothing performance deviation data stream, until the residual component of the long-term smoothing performance deviation data stream is a monotonic function or the number of extreme points of the residual component is less than a preset value.

9. The centrifugal pump performance evaluation method based on cloud computing according to claim 7, characterized in that, The step of performing energy analysis on each decomposed frequency component to identify frequency components with significant energy peaks includes: For each decomposed frequency component, calculate its energy distribution within a preset time window to obtain the energy distribution curve; Identify the energy peak point in the energy distribution curve; For each energy peak point, calculate its peak width and peak height; Based on the peak width and peak height, and combined with a preset noise characteristic threshold, it is determined whether the energy peak point satisfies the true periodic fluctuation. When the energy peak point satisfies the characteristics of true periodic fluctuation, the frequency component is identified as a frequency component with a significant energy peak.

10. A cloud computing-based centrifugal pump performance evaluation system, characterized in that, The system includes: The power calculation module is used to calculate the actual motor input power, actual head, and actual efficiency based on the centrifugal pump's operating data. The operating data includes the centrifugal pump's inlet pressure, outlet pressure, liquid flow rate, motor input current, and motor input voltage. The benchmark establishment module is used to obtain the theoretical clean water head and theoretical clean water efficiency corresponding to the liquid flow rate based on the preset clean water benchmark performance data, and to calculate the theoretical clean water motor input power based on the theoretical clean water head and the theoretical clean water efficiency. The resistance influence calculation module is used to calculate the fluid resistance influence based on the actual motor input power and the theoretical clean water motor input power. The standard correction module is used to correct the theoretical clean water head and theoretical clean water efficiency based on the fluid resistance influence amount, and generate the target theoretical clean water head and target theoretical clean water efficiency. The performance diagnostic module is used to compare the actual head and actual efficiency with the target theoretical clean water head and target theoretical clean water efficiency to obtain the comparison results. Based on the comparison results and the changing trend of the fluid resistance influence, it distinguishes between performance fluctuations caused by changes in slurry characteristics and performance degradation caused by physical wear of the pump body.

Images